<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic What Is – Agent mode in Genie? (Quick answers vs investigations) in Generative AI</title>
    <link>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/170776#M2156</link>
    <description>&lt;P&gt;&lt;STRONG&gt;Imagine asking a colleague two questions on the same morning.&lt;/STRONG&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;The first: &lt;/SPAN&gt;&lt;STRONG&gt;"What was our travel spend last month?"&lt;/STRONG&gt;&lt;SPAN&gt; They open a report, read off a number, and you're done in a minute. That's a lookup.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;The second: &lt;/SPAN&gt;&lt;STRONG&gt;"Why is our travel spend so much higher than last year?"&lt;/STRONG&gt;&lt;SPAN&gt; A good colleague won't answer that with one number. They'll think of a few possible reasons, such as more trips, pricier trips or one team travelling more, check each against the data, and come back with a short write-up and the evidence. That's an investigation.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;Genie, the Databricks tool that lets you ask questions about your data in plain English, can now work both ways. The second way is called &lt;/SPAN&gt;&lt;STRONG&gt;Agent mode&lt;/STRONG&gt;&lt;SPAN&gt;.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;The two ways Genie can answer&lt;/SPAN&gt;&lt;/H2&gt;
&lt;P&gt;&lt;STRONG&gt;Chat mode&lt;/STRONG&gt;&lt;SPAN&gt; is the quick-answer colleague. Genie writes one query and shows you the result as a table or chart. It's best for &lt;STRONG&gt;"what"&lt;/STRONG&gt;, &lt;STRONG&gt;"how much"&lt;/STRONG&gt; and &lt;STRONG&gt;"how many"&lt;/STRONG&gt; questions.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;STRONG&gt;Agent mode&lt;/STRONG&gt;&lt;SPAN&gt; is the investigating colleague who asks "&lt;STRONG&gt;why"&lt;/STRONG&gt;. Genie plans its approach, runs several queries one after another (each informed by the last), and writes up what it found in a short report. You can follow every step and the query behind it, so you can check its reasoning instead of just trusting the conclusion.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;A real-world example&lt;/SPAN&gt;&lt;/H2&gt;
&lt;P&gt;&lt;SPAN&gt;A finance team I worked with kept facing the same question: &lt;/SPAN&gt;&lt;STRONG&gt;&lt;I&gt;"Why are we over budget this quarter?"&lt;/I&gt;&lt;/STRONG&gt;&lt;SPAN&gt; Answering it meant an analyst pulling several reports into a spreadsheet and writing an email, which took days.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;With Agent mode, a manager can ask directly. Genie breaks the overspend down by department, then by supplier, spots that most of it comes from two suppliers, and writes it up. The analyst's job shifts from building the answer to checking it.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;When to use which&lt;/SPAN&gt;&lt;/H2&gt;
&lt;UL&gt;
&lt;LI&gt;&lt;SPAN&gt;&lt;STRONG&gt;Use chat mode&lt;/STRONG&gt; when you know &lt;STRONG&gt;exactly what you want: a number, a list, a chart. &lt;/STRONG&gt;&lt;/SPAN&gt;&lt;/LI&gt;
&lt;LI&gt;&lt;SPAN&gt;&lt;STRONG&gt;Use Agent mode&lt;/STRONG&gt; for "&lt;STRONG&gt;why" &lt;/STRONG&gt;and&lt;STRONG&gt; "what's driving" questions&lt;/STRONG&gt;. Agent mode is slower because it does more work, so it's the wrong choice for simple lookups.&lt;/SPAN&gt;&lt;/LI&gt;
&lt;/UL&gt;
&lt;P&gt;&lt;SPAN&gt;One caution: an investigation is only as good as the briefing. If Genie hasn't been told how your team defines "spend" or "budget", you'll get a confident report about the wrong thing. Read the steps, not just the conclusion.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;What's the first "why" question you'd hand to an investigating colleague?&lt;/SPAN&gt;&lt;/P&gt;</description>
    <pubDate>Tue, 06 Oct 2026 16:00:52 GMT</pubDate>
    <dc:creator>Valeria_Koz_DBX</dc:creator>
    <dc:date>2026-10-06T16:00:52Z</dc:date>
    <item>
      <title>What Is – Agent mode in Genie? (Quick answers vs investigations)</title>
      <link>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/170776#M2156</link>
      <description>&lt;P&gt;&lt;STRONG&gt;Imagine asking a colleague two questions on the same morning.&lt;/STRONG&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;The first: &lt;/SPAN&gt;&lt;STRONG&gt;"What was our travel spend last month?"&lt;/STRONG&gt;&lt;SPAN&gt; They open a report, read off a number, and you're done in a minute. That's a lookup.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;The second: &lt;/SPAN&gt;&lt;STRONG&gt;"Why is our travel spend so much higher than last year?"&lt;/STRONG&gt;&lt;SPAN&gt; A good colleague won't answer that with one number. They'll think of a few possible reasons, such as more trips, pricier trips or one team travelling more, check each against the data, and come back with a short write-up and the evidence. That's an investigation.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;Genie, the Databricks tool that lets you ask questions about your data in plain English, can now work both ways. The second way is called &lt;/SPAN&gt;&lt;STRONG&gt;Agent mode&lt;/STRONG&gt;&lt;SPAN&gt;.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;The two ways Genie can answer&lt;/SPAN&gt;&lt;/H2&gt;
&lt;P&gt;&lt;STRONG&gt;Chat mode&lt;/STRONG&gt;&lt;SPAN&gt; is the quick-answer colleague. Genie writes one query and shows you the result as a table or chart. It's best for &lt;STRONG&gt;"what"&lt;/STRONG&gt;, &lt;STRONG&gt;"how much"&lt;/STRONG&gt; and &lt;STRONG&gt;"how many"&lt;/STRONG&gt; questions.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;STRONG&gt;Agent mode&lt;/STRONG&gt;&lt;SPAN&gt; is the investigating colleague who asks "&lt;STRONG&gt;why"&lt;/STRONG&gt;. Genie plans its approach, runs several queries one after another (each informed by the last), and writes up what it found in a short report. You can follow every step and the query behind it, so you can check its reasoning instead of just trusting the conclusion.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;A real-world example&lt;/SPAN&gt;&lt;/H2&gt;
&lt;P&gt;&lt;SPAN&gt;A finance team I worked with kept facing the same question: &lt;/SPAN&gt;&lt;STRONG&gt;&lt;I&gt;"Why are we over budget this quarter?"&lt;/I&gt;&lt;/STRONG&gt;&lt;SPAN&gt; Answering it meant an analyst pulling several reports into a spreadsheet and writing an email, which took days.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;With Agent mode, a manager can ask directly. Genie breaks the overspend down by department, then by supplier, spots that most of it comes from two suppliers, and writes it up. The analyst's job shifts from building the answer to checking it.&lt;/SPAN&gt;&lt;/P&gt;
&lt;H2&gt;&lt;SPAN&gt;When to use which&lt;/SPAN&gt;&lt;/H2&gt;
&lt;UL&gt;
&lt;LI&gt;&lt;SPAN&gt;&lt;STRONG&gt;Use chat mode&lt;/STRONG&gt; when you know &lt;STRONG&gt;exactly what you want: a number, a list, a chart. &lt;/STRONG&gt;&lt;/SPAN&gt;&lt;/LI&gt;
&lt;LI&gt;&lt;SPAN&gt;&lt;STRONG&gt;Use Agent mode&lt;/STRONG&gt; for "&lt;STRONG&gt;why" &lt;/STRONG&gt;and&lt;STRONG&gt; "what's driving" questions&lt;/STRONG&gt;. Agent mode is slower because it does more work, so it's the wrong choice for simple lookups.&lt;/SPAN&gt;&lt;/LI&gt;
&lt;/UL&gt;
&lt;P&gt;&lt;SPAN&gt;One caution: an investigation is only as good as the briefing. If Genie hasn't been told how your team defines "spend" or "budget", you'll get a confident report about the wrong thing. Read the steps, not just the conclusion.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;What's the first "why" question you'd hand to an investigating colleague?&lt;/SPAN&gt;&lt;/P&gt;</description>
      <pubDate>Tue, 06 Oct 2026 16:00:52 GMT</pubDate>
      <guid>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/170776#M2156</guid>
      <dc:creator>Valeria_Koz_DBX</dc:creator>
      <dc:date>2026-10-06T16:00:52Z</dc:date>
    </item>
    <item>
      <title>Re: What Is – Agent mode in Genie? (Quick answers vs investigations)</title>
      <link>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/174287#M2239</link>
      <description>&lt;P&gt;If we start with a chat mode and need to use agent mode in context or continuation to the previous question, is it necessary to start a new chat with agent mode?&lt;/P&gt;</description>
      <pubDate>Thu, 22 Oct 2026 10:20:09 GMT</pubDate>
      <guid>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/174287#M2239</guid>
      <dc:creator>vuj01</dc:creator>
      <dc:date>2026-10-22T10:20:09Z</dc:date>
    </item>
    <item>
      <title>Re: What Is – Agent mode in Genie? (Quick answers vs investigations)</title>
      <link>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/174296#M2240</link>
      <description>&lt;P&gt;Good question,&amp;nbsp;&lt;a href="https://community.databricks.com/t5/user/viewprofilepage/user-id/267748"&gt;@vuj01&lt;/a&gt;&amp;nbsp;&lt;BR /&gt;As far as I can tell,&lt;STRONG&gt; unfortunately, yes,&lt;/STRONG&gt; you'll need to start a new chat. Each Genie conversation is set &lt;STRONG&gt;as either chat or agent mode when it starts. Context also stays inside one chat, so the new one won't remember the earlier thread.&lt;/STRONG&gt;&lt;/P&gt;
&lt;P&gt;&lt;STRONG&gt;A trick/workaround here: &lt;BR /&gt;1. Start a New chat and turn on Agent mode before sending the first message.&lt;BR /&gt;2. Paste in the context from the earlier chat: &lt;/STRONG&gt;the original question, key filters and definitions, and ideally the SQL Genie generated, so the agent starts from the same logic.&lt;STRONG&gt;&lt;BR /&gt;&lt;/STRONG&gt;&lt;/P&gt;
&lt;P&gt;If you mainly picked chat mode for speed, it's worth trying Flash (Beta). It's an option you turn on when you start an Agent mode conversation, and it answers faster while keeping the agent features. Like Agent mode, it can only be switched on before the first message.&lt;/P&gt;
&lt;P&gt;---&lt;/P&gt;
&lt;P&gt;I left out the internal points about chat mode being phased out. "As far as I can tell" covers the fact that the docs don't say outright you can't switch an existing thread in the UI. If you test it first and it confirms, you could drop that hedge.&lt;/P&gt;</description>
      <pubDate>Thu, 22 Oct 2026 11:18:27 GMT</pubDate>
      <guid>https://community.databricks.com/t5/generative-ai/what-is-agent-mode-in-genie-quick-answers-vs-investigations/m-p/174296#M2240</guid>
      <dc:creator>Valeria_Koz_DBX</dc:creator>
      <dc:date>2026-10-22T11:18:27Z</dc:date>
    </item>
  </channel>
</rss>

